ArticlePDF Available

The Long-Term Costs of Family Trajectories: Women’s Later-Life Employment and Earnings Across Europe

Authors:

Abstract and Figures

The “motherhood earnings penalty” is a well-established finding in many Western countries. However, a divide between mothers and nonmothers might oversimplify reality given that the family life course has diversified over the last decades. In addition, whether family choices have consequences for women’s employment and earnings in later life is not well known, particularly in a comparative perspective. Using data on 50- to 59-year-old women from the Generations and Gender Programme, the British Household Panel Survey, and SHARELIFE for 22 European countries, we derive a typology of women’s family trajectories and estimate its association with women’s later-life employment and earnings. Whereas family trajectory–related differences with regard to employment were relatively small, our findings reveal a clear, long-lasting family trajectory gradient in earnings. Childless women (with or without a partner) as well as single mothers had higher personal earnings than women whose family trajectories combined parenthood and partnership. Moreover, in societies in which reconciliation of work and family during midlife is less burdensome, labor market outcomes of women following different family trajectories converge. Our findings show that women’s fertility and partnership behavior are inevitably interrelated and jointly influence employment and earning patterns until later in life. The results imply that promoting equal employment opportunities could have long-lasting effects on women’s economic independence.
Content may be subject to copyright.
The Long-Term Costs of Family Trajectories: Womens
Later-Life Employment and Earnings Across Europe
Joanne S. Muller
1,2
&Nicole Hiekel
1,3
&Aart C. Liefbroer
1,4,5
#The Author(s) 2020
Abstract
The motherhood earnings penaltyis a well-established finding in many West-
ern countries. However, a divide between mothers and nonmothers might over-
simplify reality given that the family life course has diversified over the last
decades. In addition, whether family choices have consequences for womens
employment and earnings in later life is not well known, particularly in a
comparative perspective. Using data on 50- to 59-year-old women from the
Generations and Gender Programme, the British Household Panel Survey, and
SHARELIFE for 22 European countries, we derive a typology of womens
family trajectories and estimate its association with womens later-life employ-
ment and earnings. Whereas family trajectoryrelated differences with regard to
employment were relatively small, our findings reveal a clear, long-lasting family
trajectory gradient in earnings. Childless women (with or without a partner) as
well as single mothers had higher personal earnings than women whose family
trajectories combined parenthood and partnership. Moreover, in societies in
which reconciliation of work and family during midlife is less burdensome, labor
market outcomes of women following different family trajectories converge. Our
findings show that womens fertility and partnership behavior are inevitably
interrelated and jointly influence employment and earning patterns until later in
life. The results imply that promoting equal employment opportunities could
have long-lasting effects on womens economic independence.
Keywords Life course .Later-life earnings .Later-life employment .Family trajectory.
Comparative motherhood penalty
https://doi.org/10.1007/s13524-020-00874-8
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s13524-020-
00874-8) contains supplementary material, which is available to authorized users.
*Joanne S. Muller
muller@nidi.nl
Extended author information available on the last page of the article
Demography (2020) 57:10071034
Published online: 23 April 2020
Introduction
The increase of female employment was the most significant change in labor markets
during the past century (Esping-Andersen 2009;Goldin2006). However, womens
labor market attachment and earnings remain closely related to their family role.
Mothersemployment rates and wages lag those of men and childless women, even
when work experience is controlled for. This motherhood (earnings) penaltyis a
well-established finding in many Western countries (e.g., Correll et al. 2007;Harkness
and Waldfogel 2003; Sigle-Rushton and Waldfogel 2007a). Country-comparative
research suggests that the strength of the motherhood effect on womens employment
and personal earnings is shaped by contextual factors, such as womens opportunities to
reconcile work and family (Abendroth et al. 2014; Budig et al. 2012,2016;
Cukrowska-Torzewska 2017; Gangl 2004; Gornick and Meyers 2003; Halldén et al.
2016; Harkness and Waldfogel 2003).
Building on the motherhood earnings penalty literature, we argue that diversifying
family patterns in the second half of the twenty-first century (Elzinga and Liefbroer
2007;Kiernan2004; Sobotka and Toulemon 2008) calls for a more refined analysis of
the consequences of womens family life courses for their labor market outcomes.
Because of increasing union instability and the postponement of parenthood, the
traditional family trajectory of early and lifelong marriage and rapid and repeated
childbearing is replaced by a variety of emerging family trajectories. A simple distinc-
tion between mothers and childless women therefore no longer reflects reality.
Furthermore, the literature on motherhood and labor market outcomes tends to focus
on womens prime years of childrearing (i.e., ages 2545). However, young mothers
decisions to quit their job or reduce their working hours not only lower their current
income but also compromise their future job prospects and wages (Davies and Joshi
1994; Killewald and Zhuo 2015; Sigle-Rushton and Waldfogel 2007b). In aging
societies, later-life labor market activity is of growing importance. In particular, it is
important to examine womenspersonal earnings because these reflect economic
independence and therefore reduced vulnerability, especially at later ages. Also,
womens personal economic activity indicates the (under)use of productive potential.
The purpose of the present study is to advance our understanding of the association
between womens family trajectories and later-life labor market outcomes. We answer
two research questions. First, how are different family trajectories associated with
womens later-life employment on the one hand and personal earnings on the other?
Second, to what extent does the association between womens family trajectories and
later-life labor market outcomes vary across countries?
This study makes three contributions to the literature. First, we apply a life course
perspective to womens family trajectory by capturing the occurrence, order, and timing
of family events as a coherent chain of events rather than as separate incidents (Elder
et al. 2004). The literature has mainly focused on specific elements of the family life
course, such as the age at first birth or the occurrence of a divorce (Abendroth et al.
2014; Budig and England 2001; Gough and Noonan 2013;Miller2011; Pienta 1999;
however, a recent exception is Jalovaara and Fasang 2019). We show that the conse-
quences of the fertility and partnership trajectories for womens later-life labor market
outcomes can be more fully understood by acknowledging their interplay and thus
studying the family life course holistically.
J.S. Muller et al.
1008
Second, we contribute to the literature by assessing long-term consequences of family life
decisions. Labor market outcomes after age 50 are strongly influenced by cumulative
experience over the life course (Dannefer 2003; DiPrete and Eirich 2006; Mincer and
Polachek 1974). The family life course may have lifelong imprint effects on employment
and earnings. Alternatively, aging might level out the inequalities between women following
different family life courses. For instance, after children leave the parental home, the care-
burden difference between mothers and childless women diminishes, possibly leading
mothers to reenter the labor market and close part of the earnings gap.
Third, we examine the link between family trajectories and womens later-life labor
market outcomes from a comparative perspective. We investigate whether the level of
female labor force participation during womens family formation years moderated the
association between their family trajectory and later-life labor market outcomes. Our
data cover 22 countries, representing all European regions. This is the first study into
this topic with such a broad range of countries. We combine microlevel data from three
major longitudinal surveys: the Generations and Gender Programme, the British
Household Panel Survey, and SHARELIFE. All three surveys contain extensive
retrospective information on fertility and partnership trajectories and on current em-
ployment and personal earnings. Our sample consists of women aged 50 to 59 in the
early 2000s, who were born between 1943 and 1963.
Womens Family Trajectories and Later-Life Labor Market Outcomes
During the golden age of marriagein the 1950s and 1960s, a womanslifecoursewas
predictable: she would marry during her early 20s, subsequently quit the labor market,
and have multiple children. For instance, in Western Europe, the marriage barthat is,
strong or even institutionalized social normsdemanded that married women resign
from work (Festy 1980;Hakim2004). From the 1960s onward, postponement of
childbearing, union disruption, and single living became more prevalent (Billari and
Liefbroer 2010; Cherlin 2010; Hantrais and Letablier 1996; Sobotka 2004). Simulta-
neously, womens labor market participation expanded rapidly. In this section, we
discuss how the new variation in womens family trajectories associates with the
emerged inequality in womens economic activity. We start by reasoning that a
womans paid labor activities during midlife may predict her later-life employment
and earnings. Next, we explain how her family trajectory may relate to the midlife labor
market investments she is able or willing to make.
Consequences for Later-Life Economic Independence
Investmentoftimeandenergyinpaidworkduringmidlifemayrelatetolater-lifelabor
market outcomes due to human capital accumulation and path dependency. First, earnings
and employment opportunities accumulate throughout life according to acquired human
capital in the form of work experience (Mincer and Polachek 1974; Polanchek 1995).
Reduction of working hours or nonemployment during midlife therefore likely lowers
wages in later life. If women leave the labor market, their skills might become outdated,
or at least perceived as such by employers, which decreases womens opportunity to find a
well-paying, interesting, or challenging job later on (Correll et al. 2007).
The Long-Term Costs of Family Trajectories 1009
Second, womens past decisions and activities regarding work and family limit their
future options and preferences by means of established rules, habits, and selective
information (Abendroth et al. 2014; Heinz et al. 2009;Moenetal.1994;ORand and
Henretta 1982). For instance, women who devote their midlife to care work and
therefore prefer nonemployment, part-time, or flexible employment (over higher pay)
presumably continue that habit in later life, even when their children have left the
parental home. Also, if these women have ambitions to revive, boost, or change their
career, they face increased transaction costs. For instance, it takes time, effort, and
confidence to acquire new skills and socialize into an unfamiliar work setting.
Four Reasons Why Partnered Motherhood Could Stall Career Investment
If labor market investment during midlife is strongly related to later-life economic
outcomes, the key question is, Which family trajectories increase chances of having a
paid job? Here, we start by arguing that women with the most traditional family
trajectory (early childbearing and long-term partnership) may be least likely to invest
in employment during their childrearing years. In the subsequent subsection, we reason
why delayed motherhood and childlessness increase womensopportunity to earn
money, whereas the absence of a partner increases womensneed to do so. These
two mechanisms are visualized in Fig. 1.
In the cohort of women who have now entered later life (i.e., the cohort in our study), we
expect that women who spent their entire midlife as mother and partner tended to focus on
household and childcare activities, while their partners specialized in workplace activities.
Normative expectations, discrimination, financial deliberation, and self-selection could have
pulled these couples toward this gendered division of responsibilities.
First, whereas the societal role norm of motherpredominantly regards caring, the role
of fatherinvolves providing (Bielby and Bielby 1989; Blossfeld and Drobnič2001;
Myrdal and Klein 1956). Therefore, the male partner could easily establish identification
with both employment and parenting roles, while the female partner experiences a tradeoff
with these roles. Such role expectations were certainly dominant during the family-
formation years of the women in this study and may still be relevant today.
Provoked by normative expectations, discrimination may be at play. The job market
signaling model (Spence 1973) considers the hiring process as an investment under
uncertainty in which the employer applies normative preconceptions to judge the
Opportunity
Need
Childless,
with partner
Childless,
no partner
Single
mother
Partnered
mother
Fig. 1 Family trajectories and womens need and opportunity for labor market investments during their family
formation years
J.S. Muller et al.
1010
productivity of a potential employee. Women who have children may be attributed with
particular attitudes or characteristics related to their abilities and ambitions that imply
lower productivity. Partnered mothers with family-related employment discontinuities
in particular may signal less productivity to employers (Albrecht et al. 1999), which
reduces the chance of being hired or being in high-quality employment with appropri-
ate earnings. For instance, Correll et al. (2007) showed that mothers were considered
less competent and passionate in their profession than childless women.
Third, economic motivations may strengthen the preference of couples with children
for specialization. Assuming that partners pursue utility maximization, the most effi-
cient strategy would be that the partner with the highest earnings capacity focuses on
market production, while the other partner focuses on household work (Becker 1981).
The former is most likely the man, as a result of differences in mens and womens
initial investments in education and career, and because of gendered earning differences
between and within sectors. Although this specialization strategy was common in
marriage during the family formation phase of the women in this study, recent findings
have suggested that specialization within couples today occurs mostly after childbear-
ing (Juhn and Mccue 2017; Killewald and García-Manglano 2016; Killewald and
Gough 2013;Langner2015).
Finally, lower labor market investment of partnered mothers may be due to self-
selection. For example, women might react to poor employment outcomes by choosing
to invest their time in partnership and motherhood (Sigle-Rushton and Waldfogel
2007b) and have children at younger ages. Furthermore, Hakim (2000,2003)suggested
that women self-select based on preferences: for instance, women with a strong family
aspiration may prefer limited or flexible employment over higher pay.
Combining these arguments, we expect that women whose family trajectory is
dominated by motherhood and partnership have both the lowest need and the lowest
opportunity during midlife to invest in paid labor. Hence, we expect their employment
and earnings in later life to be lowest of all women. This results in our first hypothesis:
Hypothesis 1: Womens family trajectories characterized by having young children
in combination with a partnershipthat is, trajectories of partnered women who
have children early, many children, and/or large spacing between childrenare
associated with lowest employment and lowest earnings in later life.
Deviations From the Traditional Family Trajectory
In general, we expect that a family trajectory without a partner urges women to provide
for themselves and possible children. In other words, absence of a partner increases
womensneed to earn money. In addition, we expect that a family trajectory without
children increases womensopportunity to participate in the labor market because they
are less bound in their decisions to role expectations or work division strategies. In this
subsection, we discuss delayed motherhood, childless women with a partner, single
mothers, and childless women without a partner (single-living women).
Partnered womens later-life labor market outcomes may differ by timing of child-
bearing. The earlier women enter parenthood, the earlier they need to compromise on
the accumulation of human capital. The longer women experience time constraints to
invest in paid labor because of care work for small children, the longer they may need
The Long-Term Costs of Family Trajectories 1011
to stall career investments and the lower their opportunities to catch up career-wise later
in their lives. By contrast, women who delay motherhood may have not only developed
greater personal attachment to the labor market but also accumulated greater human
capital prior to parenthood (Gough 2017). This may increase their opportunities to keep
attached to the labor market during childrearing years or to reenter the labor market in
high-quality and well-paid employment after family-related employment disruption.
Based on these considerations, we formulate the second hypothesis:
Hypothesis 2: Wome n s family trajectories characterized by partnership and delayed
childbearing are associated with higher employment and higher earnings later in life,
compared with the aforementioned type of family trajectory (i.e., a partnered trajec-
tory with early childbearing, many children, and/or large spacing between children).
Next, women without children have no care burden during midlife, which increases
their opportunity to invest time and energy in labor market production. Therefore, we
would expect that these women have higher later-life employment and earnings than
mothers. Although partnered women without children have greater opportunities to
invest in their career, they may still rely to some extent on the main providing role of
their partner because of gendered norms regarding marriage or in anticipation of having
a child. We therefore formulate the following hypothesis on how the expected gradient
continues:
Hypothesis 3: Womens family trajectories characterized by partnership and no
childbearing are associated with higher employment and higher earnings later in
life, compared with the aforementioned type of family trajectories (i.e., trajectories
characterized by partnered motherhood).
Furthermore, the absence of a partner, either as a result of (repeated) union disso-
lution or lifelong singlehood, increases womensneed to provide for themselves and
their possible children. We expect a distinction between single women with children
and without children.
Along with the decreased union stability, the number of single mothers increased
(Cherlin 2010; Fokkema and Liefbroer 2008; Teachman et al. 2000). Compared
with all other family trajectories, women who spend a considerable time as a single
mother may experience the greatest need to provide. We expect that single mothers
will be the least driven by normative expectations toward motherhood: because they
simply need to care and provide simultaneously, they cannot afford to lower their
time investment in labor market activity (Roman 2017). However, single mothers
could have great difficulties reconciling work and family because they carry the
childcare burden on their own. They might be forced by their situation to take on
low-quality, precarious, low-paid, and/or low-intensity jobs (Christopher 2005;
Nieuwenhuis and Maldonado 2018), which offer less opportunity for career devel-
opment. Moreover, single mothers who reenter the labor market after a divorce
might have lowered human capital accumulation due to a period of care leave when
they had a partner and therefore have lower career prospects (Nieuwenhuis and
Maldonado 2018). In sum, although single mothers might have a high need to
provide, they have low opportunity.
J.S. Muller et al.
1012
We expect that a womans need to provide will have a stronger effect on later-life
labor market outcomes than her opportunity. Although a need might be unavoidable, an
opportunity permits freedom to choose. Therefore, we expect that single mothers (i.e.,
high need, low opportunity) will have higher later-life earnings than childless women
with a partner (i.e., high opportunity, low need). We formulate the fourth hypothesis:
Hypothesis 4: Womens family trajectories characterized by childbearing and the
absence of a partnership (single mothers) are associated with higher employment and
higher earnings later in life, compared with the aforementioned type of family
trajectories (i.e., trajectories characterized by partnered motherhood and trajectories
characterized by partnership without children).
Finally, unpartnered women without children may perceive the greatest opportunity
to invest in their professional career as their employment decisions cannot interfere
with career ambitions of their partner (Verbakel et al. 2008) nor with care obligations
toward children. Also, they have a clear need to provide for themselves. We would thus
expect these women to have the highest employment rates and earnings later in life:
Hypothesis 5: Womens family trajectories characterized by the absence of a part-
nership and by childlessness (single living) are associated with highest employment
and highest earnings later in life.
To summarize Hypotheses 15, we expect a gradient in later-life employment and
earnings, ranging from the lowest expected outcomes for women following a traditional
family trajectory of long-term partnership and early childbearing to the highest expect-
ed outcomes for unpartnered women without children.
Context Variation in the Relationship Between WomensFamily
Trajectories and Later-Life Labor Outcomes
In the previous section, we argue that womens later-life employment and earnings
depend on their opportunity and need for labor market investments during their family
formation phase. Here, we reason that this family trajectory gradient in later-life
employment outcomes depends on contextual circumstances (Elder et al. 2004;
Heinz et al. 2009). More specifically, we argue that the gradient is weaker in countries
that offered comprehensive support for maternal employment and hence where it was
common for women to remain attached to the labor market during their family
formation years.
In the 1970s and 1980s, when most women in our sample started their family life
course, female employment levels differed strongly across Europe. Normative beliefs
and policy regimes in Sweden, Denmark, and countries in eastern Europe showed the
strongest support for maternal employment. By the mid-1980s, mothers in these
countries were able to remain in full-time paid work with minimal career disruptions
(Gornick et al. 1997; Pascall and Manning 2000). Indeed, female labor force partici-
pation was highest in the Scandinavian countries (e.g., 75% in Sweden) and in the
satellite states of the Soviet Union (e.g., 78% in Estonia). However, support for
The Long-Term Costs of Family Trajectories 1013
maternal employment in other European countries, such as West Germany, the Neth-
erlands, and the United Kingdom was more limited or discontinuous (Gauthier 1996;
Gornick et al. 1997). In these countries, the childcare burden remained predominantly
on parents, resulting in more traditional gender role patterns. Female labor force
participation was moderate in western Europe and the United Kingdom, where between
43% (Belgium) and 55% (France) of women participated in the labor market. It was
lowest in the southern European countries (e.g., 33% in Spain), where less than half of
all women were in paid labor (all percentages from Generations and Gender
Programme 2016).
Womens opportunity to remain attached to the labor market during their family
formation phase depends on contextual circumstances that likely enhanced female
employment, such as progressive gender role attitudes and formalization of the care
sector. First, womens labor market participation depends on cultural norms about
gender roles and gender equality as well as actual gendered practices within households
(Fortin 2005). More gender-egalitarian norms may translate into a less gendered
division of paid and unpaid work, with men doing a greater share of housework and
care work and thereby increasing their female partners opportunities to invest in her
career.
Second, societies differ in the extent to which care for children is defamilialized:that
is, regarded as a public responsibility and provided by formal (i.e., paid or taxed) care
services (Bettio and Plantenga 2004;Esping-Andersen1999;SaracenoandKeck2010;
Tijdens 1993). The availability of such formal services is generally viewed as a
precondition for womens capacity to participate in the labor market (i.e., commodify
themselves) and have a continuous career regardless of their family choices (Akgunduz
and Plantenga 2018;Orloff1993). In addition, the affordability of care services is
critical. Especially for single mothers, the price of childcare could be too high com-
pared with their single-income resource level (Moilanen et al. 2016;Saraceno2011).
The provision of care by public services outside the home facilitates maternal
employment by saving motherstime and energy (Mandel and Semyonov 2006;
Misra et al. 2007). Also, formalization of the care system implies an expanded care
sector, with increased numbers of jobs available in (female-dominated) care work,
fostering womens employment opportunities in later life (Prince Cooke 2011). In
addition, the aforementioned cultural gender norms could amplify or even change the
effects of family policies on womens employment and earnings (Budig et al. 2012).
Mothers benefit most in terms of earnings from parental leave and childcare policies in
countries where attitudes support maternal employment.
Although public expenditures on childcare could lead to a lower midlife motherhood
occupational status penalty (Abendroth et al. 2014), women in countries with extensive
family provisions tend paradoxically to work more in female-typed occupations and
hold fewer managerial positions (Mandel and Semyonov 2006). Moreover, not all
family policies enhance womens labor market opportunities (Misra et al. 2007). Most
strikingly, extensive care leave provisions, called supported familialism by Saraceno
and Keck (2011), could demotivate women to reenter the labor market after childbirth.
Also, extensive care leave provisions may cause discrimination by employers against
young women (Mandel and Semyonov 2006) because they expect women to have
lower levels of productivity, on average. Employers could therefore be reluctant to hire
young women and/or invest in womens on-the-job training.
J.S. Muller et al.
1014
The previous discussion makes clear that women face considerable challenges
in the labor market even in societies with extensive family policies. Still, we
generally expect that in countries that facilitated the reconciliation of work and
family activities, women have had greater opportunities to invest in their work-
ing careers independently from their family life course, leading to a weaker
family trajectory gradient in later-life employment outcomes. We thus formulate
our final hypothesis:
&Hypothesis 6: Family trajectoryrelated differences in womens later-life employ-
ment and earnings are expected to be smaller in countries with higher female labor
force participation during womens family formation phase.
Methods
Data and Sample
Our study uses data from the first wave of the Generations and Gender Programme
(Fokkema et al. 2016; Vikat et al. 2007), the fifteenth wave of the British Household
Panel Survey (Perelli-Harris et al. 2015; University of Essex Institute for Social and
Economic Research 2010), and the third wave of SHARE (SHARELIFE) (Börsch-
Supan 2010;Schröder2011). These surveys were collected between 2004 and 2013
and contain comprehensive fertility and partnership histories as well as information
about current employment and earnings.
We restrict our sample to women aged 50 to 59 at the time of interview. Our
analytical sample consists of 18,656 women from 22 European countries (Aus-
tria, Belgium, Bulgaria, Czech Republic, Denmark, Estonia, France, Georgia,
East Germany, West Germany, Greece, Ireland, Italy, Lithuania, the Netherlands,
Norway, Poland, Romania, Spain, Sweden, Switzerland, and the United King-
dom). Table 1provides information about the year of data collection and sample
size per country.
Dependent Variables
The first dependent variable is employed (vs. not employed) at the time of interview,
measured by womens reply to the question of what their current main activity is.
Women answe r i n g in employment or self-employmentwere considered employed.
Women in the category not employedcomprised a diversity of activities, such as
being unemployed, looking after the home or family, or being retired.
The second dependent variable is personal net earnings, measured as the natural log
of annual earnings from a job or self-employment. Women were asked whether they
received earnings from a job or business during the last 12 months, how often they
received payment, and what the average net amount of payment was (i.e., the take-
home pay). By multiplying the payment amount the appropriate times (adjusted for
seasonal or otherwise not year-round work), we estimate the annual net earnings.
Earnings data are missing in approximately 20% of our sample. We impute the earnings
variable using multiple imputation (details can be found in the online appendix and
The Long-Term Costs of Family Trajectories 1015
Muller 2016). The results we present are not sensitive to the imputation of missing data
on the dependent variable. Running the models presented on a reduced sample with
nonimputed earnings (n= 9,500) yields identical results (available upon request).
Table 1 Data sources and sample sizes per country
Country Survey (year of data collection)
n
(data
source)
n
(country
total)
Austria SHARELIFE (20082009) 91 91
Belgium Generations and Gender Survey (GGS) (20082010) 675 1,077
SHARELIFE (20082009) 402
Bulgaria GGS (20042005) 878 878
Czech Republic GGS (2005) 908 1,208
SHARELIFE (20082009) 300
Denmark SHARELIFE (20082009) 375 375
Estonia GGS (20042005) 898 898
France GGS (2005) 1,063 1,407
SHARELIFE (20082009) 344
Georgia GGS (2006) 936 936
East Germany GGS (2005) 187 251
SHARELIFE (20082009) 64
West Germany GGS (2005) 675 886
SHARELIFE (20082009) 211
Greece SHARELIFE (20082009) 525 525
Ireland SHARELIFE (2007) 145 145
Italy SHARELIFE (20082009) 299 299
Lithuania GGS (2006) 746 746
Netherlands GGS (20022004) 871 1,200
SHARELIFE (20082009) 329
Norway GGS (20072008) 1,354 1,354
Poland GGS (20102011) 2,403 2,782
SHARELIFE (20082009) 379
Romania GGS (2005) 1,233 1,233
Spain SHARELIFE (20082009) 253 253
Sweden GGS (20122013) 838 1,028
SHARELIFE (20082009) 190
Switzerland SHARELIFE (20082009) 206 206
United
Kingdom
British Household Panel Survey (BHPS) Wave 15
(20052006)
878 878
Total GGS 13,665 18,656
SHARELIFE 4,113
BHPS 878
J.S. Muller et al.
1016
We transform the earnings measure from national currencies to international com-
parable euros in three steps. First, we convert to the year 2008 (i.e., because most data
were collected in 2008) by using the consumer price index to correct for inflation
(World Bank n.d.-a). Next, we convert these national 2008 currencies to international
comparable dollars using the purchasing power parity (PPP) conversion factor (World
Bank n.d.-b). Last, we convert to PPP euros using the annual average exchange rate in
2008 between dollars and euros (i.e., 1 euro = 1.4708 U.S. dollars; Eurostat n.d.). Using
PPPs allows us to make a more meaningful comparison between countries because it
adjusts for differences in the cost of living (World Bank 2013). Logging ensures that
the earnings distribution meets the assumption of normality and outlier effects are
minimized. Multiplying the coefficients by 100 × (eb1) gives the percentage change
in earnings, given a one-unit increase in the independent variable.
Independent Variables
At the micro level, the variable of interest is womens family trajectory, which is measured as
a sequence of yearly states from ages 18 to 50that is, 33 chronological states. We specify
eight possible states based on a combination of the age of the youngest child and partnership
status: (1) no child, no partner; (2) no child, with partner; (3) youngest child aged 03, no
partner; (4) youngest child aged 03, with partner; (5) youngest child aged 411, no partner;
(6) youngest child aged 411, with partner; (7) youngest child aged 12+, no partner; and (8)
youngest child aged 12+, with partner.
We consider the age of the youngest child because the care burden for that child is
highest. We distinguish pre-primary, primary, and post-primary school age. We include
biological and adoptive children, but we exclude stepchildren. Our assumption is that
children for which mothers have prime responsibility have most impact on their
employment need and opportunity.
The states with partnercover cohabiting and married partnerships with a male or
female partner. We include unmarried coresident couples because cohabitation is
viewed increasingly as an alternative or prelude to marriage (Hiekel et al. 2014). Also,
we regard the situation of having children in a cohabiting union closer to having
children within marriage instead of having children without a coresident partner.
At the macro level, we would ideally want to include indicators that explain country
variation in womens employment opportunities during their family formation phase
(the 1970s and 1980s for women in our data). Important variables could be measures of
gender role norms, the availability and affordability of formal childcare, the size of the
public sector, and flexible labor market opportunities. Unfortunately, such detailed
country-level data are not available, or are barely available, for the relevant period
(the 1970s and 1980s), especially for the eastern European countries. Therefore, we
focus on one general indicator that is available and reflects which countries were, at the
time, forerunners in womens (especially mothers) employment and offered relatively
comprehensive support for reconciliation of paid and unpaid work: the female labor
force participation rate. The female labor force participation rate indicates the percent-
age of the female population aged 1564 that is active on the labor market. We use
1980 data from the GGP Contextual Database (Generations and Gender Programme
2016), which is the first time point available. No macro data were available for East
Germany.
The Long-Term Costs of Family Trajectories 1017
Control Variables
First, educational attainment positively affects employment, working hours, hourly
wages, and ultimately earnings (Mincer 1975). Moreover, low education is associated
with experiencing potentially disadvantageous events in the family trajectory, such as
early parenthood and divorce (Härkönen and Dronkers 2006; Raymo et al. 2015). We
use the International Standard Classification of Education (ISCED) to distinguish three
levels of educational attainment: no or primary education (ISCED 0, 1, or 2), lower and
upper secondary education (ISCED 3 or 4), and all types of tertiary education (ISCED 5
or higher).
Second, age relates to labor market outcomes resulting from human capital accu-
mulation and selection effects (Murphy and Welch 1990). Also, age might relate to the
prevalence of certain family trajectories because of cohort effects. We measure age in
years at the time of interview.
Third, to disentangle retrospective and current family effects on labor market
outcomes, we include two variables. One indicates whether the respondent had a
coresident child younger than 18 years at the time of interview, and the other indicates
whether she had a coresident partner at the time of interview.
Finally, given that earnings depend on the combination of the number of working
hours and hourly wages, we include a dummy variable indicating whether the woman
is employed 30 hours or more per week.Wecannotuseamoredetailedmeasure
because the SHARELIFE data contain only this crude measure.
Descriptive statistics for all variables can be found in Table 2.
Analytical Approach
Our analytical approach consists of two parts. First, we create a typology of
family trajectories using sequence cluster analysis. Second, we use this typology
in regression models to predict variance in womens later-life employment and
earnings.
The strength of sequence analysis is that it provides a holistic view of trajectories,
which allows us to determine trajectory patterns, taking into account ordering (se-
quencing), timing of family events, and duration of states (Cornwell 2015; Studer and
Ritschard 2014). In sequence analysis, similarities between trajectories are expressed as
distancesthat is, trajectories that strongly resemble each other have a short distance to
each other, whereas trajectories that are very different have a large distance. First, we
calculate a pairwise distance matrix by using optimal matching with a constant
substitution matrix and indel cost of 1 (Studer and Ritschard 2014). Next, we perform
a hierarchical cluster analysis on this distance matrix using Wards method, which
implies that sequences with the smallest distance from each other are clustered. To
determine the most appropriate number of clusters, we consider two cluster cutoff
criterianamely, the average silhouette widths (ASW) and point biserial correlation
(PBC) (Studer 2013)as well as the construct validity of the cluster solution, by
comparing the fit of regressions of the different cluster-solutions on earnings
Akaike information criterion (AIC) and Bayesian information criterion (BIC) (Han
et al. 2017; Warren et al. 2015). We use the TraMineR package in R to perform the
sequence cluster analysis (Gabadinho et al. 2011).
J.S. Muller et al.
1018
Table 2 Descriptive information per variable per country
Educational Level Family Trajectory Cluster
Var i a b l e
Age
(range =
5059
years)
Birth Year
(range =
19431963)
1=
Low
2=
Middle
3=
High
Coresident
Children
(1 = 1+
children)
Coresident
Partner
(1 = yes)
Child
With
Partner,
Stretched
Child
With
Partner,
Early
Child
With
Partner,
Delayed
Single
Mother
No
Child,
With
Partner
No
Child,
No
Partner
Employed
(1 = yes)
Wor k s
Full-
Timea
(1 =
yes)
Log
(Earnings)a
(range =
(0,])
Austria 55.5 1953 30.8 55.0 14.3 6.6 87.9 25.3 39.6 14.3 11.0 2.2 7.7 39.6 31.9 9.3
Belgium 54.7 1954 36.6 32.5 30.9 10.6 76.3 16.3 35.4 16.0 17.9 9.8 4.6 58.6 30.9 9.4
Bulgaria 54.6 1949 27.7 48.0 24.4 19.5 78.6 13.9 54.4 11.3 12.9 4.2 3.3 56.6 45.4 7.9
Czech
Republic
54.8 1951 27.7 61.9 10.4 8.3 65.2 13.6 40.6 9.9 21.6 7.3 7.0 63.5 54.1 8.6
Denmark 54.9 1954 12.0 33.1 54.9 8.3 83.2 22.1 29.9 21.6 14.4 7.5 4.5 83.7 61.3 9.6
Estonia 54.3 1950 13.7 49.1 37.2 10.9 65.6 25.2 34.3 7.5 26.0 3.7 3.5 74.5 69.3 8.6
France 54.7 1951 39.5 37.2 23.3 11.6 66.9 20.1 32.5 13.4 21.7 5.2 7.1 65.0 43.4 9.3
Georgia 54.2 1952 10.2 62.0 27.9 37.2 69.2 19.1 45.0 11.9 13.6 1.5 9.0 49.4 27.1 6.9
East
Germany
54.6 1951 18.7 62.6 18.7 8.0 74.1 19.9 43.0 10.0 14.3 4.8 8.0 72.1 48.6 9.2
Wes t
Germany
54.5 1951 34.7 50.1 15.2 13.7 74.3 16.0 27.3 19.8 17.2 12.6 7.1 70.2 35.9 9.3
Greece 55.2 1954 37.7 42.3 20.0 4.8 81.9 18.7 48.4 12.8 8.2 6.5 5.5 36.8 35.1 9.3
Ireland 54.9 1952 24.8 26.9 48.3 18.6 75.2 40.0 16.6 17.2 13.8 4.8 7.6 46.9 30.3 9.3
Italy 55.1 1954 55.9 35.8 8.4 10.7 88.0 27.8 44.2 11.4 6.0 5.7 5.0 43.1 33.8 9.3
Lithuania 54.2 1952 10.3 63.3 26.4 15.6 52.4 15.8 29.1 11.8 27.8 5.5 10.1 72.4 60.1 9.6
Netherlands 54.5 1950 48.0 24.8 27.3 9.2 72.3 15.2 33.3 17.0 16.5 9.8 8.3 58.1 22.1 9.2
Norway 54.4 1953 16.0 48.2 35.8 18.1 72.5 27.4 28.6 12.9 21.3 4.9 5.0 88.6 56.3 9.8
Poland 54.9 1955 18.5 67.5 14.0 16.7 67.3 26.3 38.5 7.4 17.7 4.1 6.1 42.9 34.9 8.7
Romania 54.4 1951 52.3 39.2 8.5 17.5 73.6 19.2 45.2 10.5 14.5 8.3 2.4 41.4 28.5 7.7
The Long-Term Costs of Family Trajectories 1019
Table 2 (continued)
Educational Level Family Trajectory Cluster
Var i a b l e
Age
(range =
5059
years)
Birth Year
(range =
19431963)
1=
Low
2=
Middle
3=
High
Coresident
Children
(1 = 1+
children)
Coresident
Partner
(1 = yes)
Child
With
Partner,
Stretched
Child
With
Partner,
Early
Child
With
Partner,
Delayed
Single
Mother
No
Child,
With
Partner
No
Child,
No
Partner
Employed
(1 = yes)
Wor k s
Full-
Timea
(1 =
yes)
Log
(Earnings)a
(range =
(0,])
Spain 55.2 1953 67.6 21.0 11.5 9.1 85.0 28.5 41.1 10.3 4.0 8.3 7.9 45.5 39.5 9.3
Sweden 54.7 1957 11.0 48.2 40.9 16.5 75.0 21.8 22.4 23.7 21.8 5.0 5.4 92.5 71.5 9.8
Switzerland 55.2 1953 26.2 62.6 11.2 6.8 74.8 16.5 34.5 16.0 16.0 5.8 11.2 69.9 23.8 9.5
United
Kingdom
54.4 1950 28.1 39.4 32.4 8.2 69.7 19.8 29.4 15.8 20.5 5.6 9.0 66.0 37.0 9.4
Total 54.6 1952 27.8 48.3 23.9 14.4 71.2 20.5 36.1 13.0 18.1 6.1 6.2 61.2 42.4 9.0
aValues represent the percentage of full-time workers and average earnings among employed women.
J.S. Muller et al.
1020
In the second step of our analysis, we use the family trajectory typology that results
from the sequence cluster analysis to predict womens later-life likelihood of being
employed and the earnings of employed women.
To answer our first research question on how later-life employment and earnings
systematically vary by womens midlife family trajectory, we estimate pooled logistic
(employment) and linear (earnings) models with country fixed effects. We include the
family trajectory typology as a categorical variable (i.e., a set of dummy variables) to
these models. To test Hypotheses 15, we contrast different family trajectory clusters
with each other (by simply switching the reference category of the set of dummy
variables). In these models, we add a variance-covariance matrix (VCE) cluster cor-
rection that adjusts standard errors for intracountry correlation (Huber 1967; Rogers
1993;White1980). Because we are not interested in absolute income differences
between countries, we center the dependent earnings variable by country data set; that
is, we subtract the country-specific mean log(earnings) from the log(earnings) variable.
To answer our second research question regarding the extent to which the associa-
tion between womens family trajectories and later-life labor market outcomes varies
across countries, we include interaction effects between the family trajectory typology
dummy variables and country dummy variables in the regression models. Using a chi-
square test for employment and an analysis of variance (ANOVA) for earnings,
respectively, we test the interaction of the two categorical variables: country and family
trajectory typology. This is an appropriate method to assess country effects in a study
such as ours with relatively few Level 2 cases, i.e. only 22 countries (Bryan and Jenkins
2016; Cameron and Miller 2015; Snijders and Bosker 2012).
Finally, to examine whether the family trajectory gradient in later-life employment
and earnings is moderated by the midlife level of female labor force participation in a
given country, we test interaction effects between the family trajectory clusters and the
1980 macro-level female labor force participation rate.
Results
Typology of Family Trajectories
The cluster analysis of the sequence distance matrix results in a six-cluster solution.
Although the two cluster cutoff criteriaASW and PBCindicate that the four-cluster
solution would be optimal, the regression fit indicesAIC and BICare superior for
the six-cluster solution (Table 3). We choose the six-cluster solution because in the
four-cluster solution, 69.6% of women were assigned to one cluster; comparatively, in
the six-cluster solution, 36.1% of women were in the largest cluster. Moreover,
compared with the four-cluster solution, the six-cluster solution grasps more relevant
detail of the family trajectory.
Figure 2shows the sequence index plots of the six family trajectory clusters. We
label each cluster based on its characteristics. We identify two types of traditional
motherhood trajectories: child with partner, stretched (CWP stretched)andchild with
partner, early (CWP early). Women in these two clusters experienced the same
sequence: they started living with a partner early in their adult life, had one or more
children, and stayed together with their partner until at least age 50. The difference
The Long-Term Costs of Family Trajectories 1021
between the two clusters is the number and spacing of children. CWP stretched implies
that women in this cluster had many children or a large time gap between births and
therefore had an extended period of care burden.
The remaining four trajectories corresponded to the discussed deviations from the
traditional, partnered motherhood trajectory. First, women in the CWP delayed cluster
delayed partnering and motherhood. Second, we identify two clusters of childless
women who spent most of their life (1) with a partner: no child, with partner (NCWP)
or (2) without a partner: no child, no partner (NCNP).A final cluster comprised women
who experienced a substantial spell of single motherhood. Some of these single
Table 3 Fit indices of several cluster solutions of family trajectories: Average silhouette widths (ASW), point
biserial correlation (PBC), Akaike information criterion (AIC), and Bayesian information criterion (BIC)
Number of Clusters ASW PBC AIC BIC
4 0.47 0.76 21,708.0 21,810.8
5 0.26 0.51 21,709.1 21,819.2
6 0.28 0.49 21,688.4 21,805.9
7 0.28 0.51 21,690.3 21,815.1
8 0.19 0.41 21,692.2 21,824.4
9 0.20 0.41 21,691.8 21,831.3
Notes: AIC and BIC are estimated in a linear regression model with centered log(earnings) as dependent
variable; independent variables are age in years, educational level, country and family trajectory, and a VCE
cluster correction (data set).
No child no partner (single living)
No child with partner (childless/childfree)
Single mother
Child with partner, early
Child with partner, stretched
Child with partner, delayed
20 25 30 35 40 45 50
No child
0–3 4–11 12+
No partner
With partner
Age of youngest child
20 25 30 35 40 45 50
20 25 30 35 40 45 50 20 25 30 35 40 45 50
20 25 30 35 40 45 50
20 25 30 35 40 45 50
Fig. 2 Sequence index plots of womens family trajectories between ages 18 and 50 across 22 European
countries. n= 18,656.
J.S. Muller et al.
1022
mothers started their trajectory traditionally; they coresided with a partner and had a
child. However, after union dissolution, they continued living with their child(ren)
without a partner. Other single mothers spent most of their life without a coresiding
partner and raised their child(ren) on their own.
Most (69.6%) women in our sample were in one of the three CWP clusters. In
addition, 12.3% of women were in one of the two childless trajectories (NCWP or
NCNP), and 18.1% of women were in the single mother cluster. Table 2shows the
distribution of family trajectory clusters in total and by country.
Estimating Employment and Earnings Differences Between Family Trajectories
Next, we include the derived family typology as a set of dummy variables in logistic
regression models estimating whether women were in paid employment and in linear
regression models estimating earnings among those women who were. Results on
employment are presented in Table 4, and results on earnings are presented in Table 5.
The first model in Table 4shows that whether women aged 5059 in Europe were in
paid employment increases with level of educational attainment and decreases with age.
To examine differences between family trajectories, we first perform a chi-square test
assessing the joint effect of all family trajectory dummy variables, which is significant
(χ2(5) = 41.61, p< .01). This implies that whether women are employed in later life
differs significantly between family trajectory types.
Next, we examine whether the results regarding employment were in line
with Hypotheses 15 by taking different family trajectories as reference cate-
gory (detailed results available upon request). This is generally not the case.
Women whose family trajectory was characterized by partnership and delayed
childbearing (i.e., the CWP delayed cluster) had the highest employment rate.
Differences between women in the other groups were relatively small and
mostly nonsignificant.
Results in Model 2 of Table 4show that the differences in employment between the
family trajectory clusters remain largely the same when current family situation (i.e.,
whether women had a partner and/or a child in the household at the time of the
interview) is controlled for.
Last, we find significant variation in the effect of family trajectory between countries
(χ2(11) = 10,857.60, p< .01). In Model 3, we add an interaction between the female
labor force participation rate in 1980 and the family trajectory clusters. Several
interaction terms as well as the combined interaction effects are statistically
significant (χ2(5) = 23.51, p< .01). To facilitate interpretation of the interaction
effect, we graph the predicted employment rate for all trajectory clusters across the
levels of female labor force participation observed in the sample of countries (Fig. 3).
Results are in line with the hypothesis. In countries with a low level of female labor
force participation in 1980, differences in paid employment at ages 5059 were
relatively large between the family trajectory clusters, with women in the CWP
stretched cluster having the lowest employment. In countries with a high level of
female labor force participation in 1980, differences in paid employment at ages 5059
between women with different family trajectories were smaller.
Tab le 5shows models predicting later-life earnings among women in paid employment.
Earnings increased with educational attainment, but we observe no clear age pattern. To
The Long-Term Costs of Family Trajectories 1023
Table 4 Logistic regression models predicting later-life employment of women
123
Educational Level (ref. = middle education)
Low education 0.66*** 0.65*** 0.59***
(0.06) (0.06) (0.12)
High education 0.86*** 0.86*** 0.99***
(0.06) (0.06) (0.10)
Age in Years 0.16*** 0.17*** 0.16***
(0.02) (0.02) (0.02)
Current Characteristics
Coresident partner (yes) 0.12*
(0.06)
Coresident child <18 (yes) 0.14**
(0.05)
Early and Midlife Family Trajectory Typology (ref. = child with partner, stretched)
Child with partner, early 0.08 0.05 0.73*
(0.05) (0.05) (0.30)
Child with partner, delayed 0.30*** 0.30*** 0.64
(0.07) (0.07) (0.40)
No child with partner 0.11 0.07 1.70***
(0.11) (0.10) (0.43)
Single mother 0.13 0.03 0.93*
(0.09) (0.09) (0.44)
No child, no partner 0.06 0.18 1.30*
(0.12) (0.12) (0.53)
Country-Level Variables and Interactions (ref. = female labor force participation
(FLFP) 1980 × child with partner, stretched)
FLFP 1980 2.37*
(0.95)
FLFP 1980 × Child with partner, early 1.22*
(0.50)
FLFP 1980 × Child with partner, delayed 0.40
(0.67)
FLFP 1980 × No child, with partner 2.59***
(0.68)
FLFP 1980 × Single mother 1.19
(0.70)
FLFP 1980 × No child, no partner 2.26*
(0.88)
Constant 9.54*** 9.91*** 7.43***
(1.18) (1.19) (1.38)
n18,656 18,656 18,405
Notes: Standard errors are shown in parentheses. The reference category for educational level is middle
education. The reference category of family trajectory typology is child with partner, stretched. VCE cluster
correction (data set) is included in all models. Country dummy variables are included in Model 1 and Model 2
(country coefficients are not shown in the table).
p<.10;*p< .05; **p<.01;***p< .001 (two-sided tests)
J.S. Muller et al.
1024
Table 5 Linear regression models predicting later-life earnings of employed women
123
Educational Level (ref. = middle education)
Low education 0.29*** 0.24*** 0.28***
(0.04) (0.03) (0.04)
High education 0.37*** 0.34*** 0.35***
(0.03) (0.04) (0.03)
Age in Years 0.01 0.00 0.01
(0.00) (0.00) (0.00)
Current Characteristics
Coresident partner (yes) 0.03
(0.03)
Coresident child <18 (yes) 0.04
(0.03)
Working hours 30+ (yes) 0.55***
(0.07)
Early and Midlife Family Trajectory Typology (ref. = child with partner, stretched)
Child with partner, early 0.06* 0.030.03
(0.02) (0.02) (0.10)
Child with partner, delayed 0.10** 0.08** 0.22
(0.02) (0.02) (0.14)
No child, with partner 0.17*** 0.10** 0.47*
(0.04) (0.03) (0.22)
Single mother 0.14*** 0.06* 0.22
(0.02) (0.02) (0.18)
No child, no partner 0.22*** 0.11** 0.73*
(0.05) (0.03) (0.27)
Country-Level Variables and Interactions (ref. = female labor force participation (FLFP) 1980 × child with partner, stretched)
FLFP 1980 0.11
(0.15)
FLFP 1980 × Child with partner, early 0.06
(0.17)
FLFP 1980 × Child with partner, delayed 0.19
(0.22)
FLFP 1980 × No child, with partner 0.48
(0.36)
FLFP 1980 × Single mother 0.13
(0.28)
FLFP 1980 × No child, no partner 0.82
(0.42)
Constant 0.21 0.25 0.33
(0.25) (0.21) (0.31)
n11,415 11,2 55 11,234
Notes: Standard errors are shown in parentheses. Sample contains employed or self-employed women only.
Dependent variable: centered log (earnings). The reference category of educational level is middle education.
The reference category of family trajectory typology is child with partner, stretched. VCE cluster correction
(data set) is included in all models. Country dummy variables are included in Model 1 and Model 2 (country
coefficients are not shown in the table).
p<.10;*p< .05; **p<.01;***p< .001 (two-sided tests)
The Long-Term Costs of Family Trajectories 1025
examine the overall effect of the family trajectory typology, we perform an ANOVA test
assessing the joint effect of all family trajectory dummy variables, which is significant
(F(5,26) = 9.33, p< .01). Subsequently, we test Hypotheses 15bytakingdifferentfamily
trajectories as reference category (again, by contrasting different trajectories against each
other; detailed results available on request). Figure 4presents the relative earnings by family
trajectory cluster, based on Model 1 in Table 5. In general, we find that the order of the
clusters corresponds with the expectations in Hypotheses 15. However, not all differences
between clusters are statistically significant.
In line with the first hypothesis, our findings indicate that being a mother with a
traditional family trajectorythat is, having children at a young age and a lifelong
partnerpenalizes women most in terms of personal earnings. Furthermore, women
with the trajectory CWP delayed earned, on average, 10.2% more in later life than
women with trajectory CWP stretched. This implies that in line with the second
0.5
0.7
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
2.5
.30 .40 .50 .60 .70 .80
Relative Odds Ratio (CWP, stretched = 1)
Female Labor Force Participation, 1980
No child, no partner Child with partner, delayed
No child with partner Child with partner, early
Single mother Child with partner, stretched
Fig. 3 Relative odds ratio to be employed for women aged 5059, by type of family trajectory and level of
female labor force participation in 1980 (child with partner, stretched = 1). Coefficients are exponentiated
(based on Table 4, Model 3). Traditional trajectories have a solid fill; deviations have no fill.
0.90
0.95
1.00
1.05
1.10
1.15
1.20
1.25
1.30
Child With
Partner,
Stretched
Child With
Partner,
Early
Child With
Partner,
Delayed
No Child
With Partner
Single
Mother
No Child,
No Partner
Relative Earnings (CWP, stretched = 1)
Fig. 4 Relative earnings of women employed at age 5059 by type of family trajectory (child with partner,
stretched = 1). Coefficients are exponentiated (based on Table 5, Model 1). Traditional trajectories have a solid
fill; deviations are striped.
J.S. Muller et al.
1026
hypothesis, among partnered mothers, the longer the time women spent with dependent
children, the lower their earnings in later life.
Next, we find that NCWP and single mothers indeed had higher earnings than all
partnered motherhood trajectories (in line with Hypotheses 3 and 4). However, single
mothers and NCWP did not significantly differ from each other, contrary to our
expectation (Hypothesis 4). Last, women with the NCNP trajectory had significantly
higher earnings than all other women, as expected (Hypothesis 5), except for women
with the NCWP trajectory, which was not expected. For instance, on average, women
with the NCNP trajectory earned 24.9% more annually than women in the CWP
stretched trajectory. Combining the insights from Hypotheses 15, we find a gradient
in womens later-life earnings according to their family trajectory as shown in Fig. 4.
Model 2 of Table 5shows that differences in earnings between trajectory clusters
become somewhat smaller only when controlled for current partner status, currently
having a child in the household, and full-time working hours. This suggests that the
differences between clusters at least partially result from differences in earning potential
between women and not just from differences in the number of working hours.
Finally, we find significant variation in the effect of family trajectory on earnings
between countries (F(43,75.7) = 15.82, p<.01). In Model 3 of Table 5,weexamine
those country differences by adding interaction effects between trajectory clusters and
the country-level female labor force participation ratio in 1980. Hypothesis 6 states that
higher female labor force participation in a given country at the time of womensfamily
formation (1980) would be associated with smaller differences in later-life earnings
between women with different family trajectories. The overall picture, visualized in
Fig. 5, is in line with this hypothesis. In countries with low levels of female labor force
participation differences in earnings between women with different family trajectories
are substantial, but they are small in countries with high levels of female labor force
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
.30 .40 .50 .60 .70 .80
Relative Earnings (CWP, stretched = 1)
Female Labor Force Participation, 1980
No child, no partner Child with partner, delayed
No child with partner Child with partner, early
Single mother Child with partner, stretched
Fig. 5 Relative earnings of women employed at age 5059 by type of family trajectory and level of female
laborforceparticipationin1980(child with partner, stretched = 1). Coefficients exponentiated (based on
Tab le 5, Model 3). Traditional trajectories have a solid fill; deviations have no fill.
The Long-Term Costs of Family Trajectories 1027
participation. However, only the difference between the NCNP cluster and the CWP
stretched cluster statistically significantly weakens. In sum, our results are in line with
Hypothesis 6, but only the difference between the two extreme clusters declines
significantly by countrys female labor force participation rate.
Discussion and Conclusion
In this study, we examine whether womens family trajectories have long-term conse-
quences for employment and personal earnings. Using data on 22 European countries,
we derive six family trajectory types based on timing and number of children, and
timing and number of partnerships between ages 18 and 50. Subsequently, we use this
typology to predict differences in employment and earnings among women aged 50
59. Our work extends previous literature on the motherhood earnings penalty by
examining the intertwined fertility and partnership trajectory, studying their long-term
effects, and taking a comparative perspective.
Our first main finding is that contrary to what prior research on the mother-
hood earnings penalty suggests (e.g., Harkness and Waldfogel 2003;Sigle-
Rushton and Waldfogel 2007a), there is no strict employment and earnings
divide between mothers and nonmothers. Rather, we find little variation in
womens later-life employment, and we find evidence for a gradient in womens
earnings based on the diversity in their family trajectories. Especially the earn-
ings of single mothers and partnered women who delayed motherhood are
similar to those of childless women. Our results indicate that an earnings penalty
exists mostly for women with the most traditional motherhood trajectory of
lifelong coresident partnership, early motherhood, and multiple children. On the
other hand, women who live without a partner and without children the majority
of their life (i.e., single living) have the highest earnings at the end of their
careers.
We find that the higher womensneed to provide during midlife (i.e., the more time
without partner) and the more opportunity to invest in the labor market (i.e., the more
time without dependent children), the higher their later-life earnings. However, we do
not find evidence that the need to provide is more important than the opportunity: the
earnings levels of single mothers (high need, low opportunity) and partnered childless
women (low need, high opportunity) do not differ.
Although we observe large earnings differences between women according to their
family trajectory, we find relatively small differences in employment. Partnered women
who delayed childbearing had the highest employment rate, suggesting that their
delayed pattern of childbearing increased their chances of remaining attached to the
labor market. However, contrary to our expectations, we find that single mothers and
childless women (with or without children) did not have higher employment rates than
partnered mothers.
The second main finding of this study is that the midlife family trajectory has
long-lasting consequences for womens personal earnings and employment be-
yond the childrearing years. Our findings of a family trajectory gradient in
womens later-life earnings, even when we take into account their current family
situation, strengthen our confidence that womens family trajectories have long-
J.S. Muller et al.
1028
lasting implications for their earnings. This result puts previous findings on
midlife motherhood earnings penalties in a broader life course perspective
(e.g., Gangl and Ziefle 2009). The accumulating nature of labor market earnings
makes it vital for women to have the opportunity to reconcile work and family
activities in order to be more economically independent later in life.
Our third main finding is that context matters. Confirming our hypothesis
regarding country variation, womens family trajectory gradient in employment
and earnings was smaller in countries with higher female labor force participa-
tion in 1980, when women in our study had to reconcile care work and career
investments. This suggests that in societies in which reconciliation of work and
family during midlife is less burdensome, labor market outcomes of women
following different family trajectories converge and hence decrease economic
inequality between women until the end of their careers.
In addition to these contributions, our study leaves a number of interesting questions
unanswered. First, an important topic for future research would be to disentangle the
effects for working hours and hourly wages. We show that working full-time hours
does not fully mediate the effect of family trajectory on later-life earnings, suggesting
that earnings of women with different trajectories vary because of both differences in
working hours and in wage per hour. Our data do not allow a more detailed analysis of
this matter, nor can we decompose the comparative findings in effects of working hours
and hourly wages.
Second, a main strength of our analytical approach is that it appreciates the full
family trajectory, the interdependence of life events, and their combined meaning. At
the same time, we cannot pinpoint whether specific aspects of these trajectories (e.g.,
occurrence, timing, and/or duration of family spells) matter. Future research could
answer such questions, needing highly detailed panel family and earnings data and
using a fixed-effects approach.
Moreover, such prospective panel data would be beneficial because these data
do not suffer from any recall bias. For the current study, we use retrospective
family trajectory data. Although people usually do not forget when their children
were born, they might not report relationships, especially those that were brief.
This study therefore potentially underestimates the number of partnerships ever
entered or dissolved. Longitudinal, career-long data are available in only a
handful of often-studied countries, however, making these data hard to apply
in a comparative perspective.
Third, we focus only on women and their personal earnings. Retrospective partner
information is lacking in our data sets, but it would be interesting to examine the role of
partner characteristics.
Last, as we indicated earlier, ideally we would have used more detailed macro
variables in our comparative analysis. This would allow us to disentangle the relevance
of, for instance, different family policies, labor market characteristics, and societal role
norms. However, more specific indicators (e.g., the availability and affordability of
formal childcare, the size of the public sector, gender inequality, gender role norms, and
flexible labor market opportunities) are barely available for the 1970s and 1980s. Most
information is available from the 2000s onward, especially in the former Soviet
countries. Our strongest recommendation for future research is, therefore, that we need
better historical data on family policies and other relevant societal aspects to further
The Long-Term Costs of Family Trajectories 1029
unravel the mechanisms behind womens divergent levels of later-life economic
resilience.
Acknowledgments The research leading to these results has received funding from the European Research
Council under the European Unions Seventh Framework Programme (FP/2007-2013) / ERC Grant Agree-
ment No. 324178 (Project: Contexts of Opportunity. PI: Aart C. Liefbroer).
AuthorsContributions All authors contributed to the study concept and design. Data harmonization and
data preparation were performed by Joanne Muller. Data analyses were performed by Joanne Muller and Aat
Liefbroer. The first draft of the manuscript was written by Joanne Muller. All authors commented on
subsequent versions of the manuscript. All authors read and approved the final manuscript.
Data Availability All data sets used are publicly available from GGP, SHARE, and BHPS. The GGP
earnings data set prepared specifically for this study is available on the GGP website. A publication package
including all syntaxes is available on request from the first author.
Compliance with ethical standards
Ethics and Consent The authors report no ethical issues.
Conflict of Interest The authors declare no conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence,and
indicate if changes were made. The images or other third party material in this article are included in the
article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not
included in the article's Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
References
Abendroth, A.-K., Huffman, M. L., & Treas, J. K. (2014). The parity penalty in life course perspective:
Motherhood and occupational status in 13 European countries. American Sociological Review, 79,993
1014.
Akgunduz, Y. E., & Plantenga, J. (2018). Child care prices and maternal employment: A meta-analysis.
Journal of Economic Surveys, 32,118133.
Albrecht, J. W., Edin, P.-A., Sundström, M., & Vroman, S. B. (1999). Career interruptions and subsequent
earnings: A reexamination using Swedish data. Journal of Human Resources, 34,294311.
Becker, G. S. (1981). A treatise on the family. Cambridge, MA: Harvard University Press.
Bettio, F., & Plantenga, J. (2004). Comparing care regimes in Europe. Feminist Economics, 10(1), 85113.
Bielby, W. T., & Bielby, D. D. (1989). Family ties: Balancing commitments to work and family in dual earner
households. American Sociological Review, 54,776789.
Billari, F. C., & Liefbroer, A. C. (2010). Towards a new pattern of transition to adulthood? Advances in Life
Course Research, 15,5975.
Blossfeld, H.-P., & Drobnič, S. (2001). Theoretical perspectives on couplescareers. In H.-P. Blossfeld & S.
Drobnič(Eds.), Careers of couples in contemporary societies: From male breadwinner to dual-earner
families (pp. 1650). Oxford, UK: Oxford University Press.
Börsch-Supan, A. (2010). Survey of Health, Ageing and Retirement in Europe (SHARE) Wave 3
SHARELIFE. Munich, Germany: SHARE-ERIC. https://doi.org/10.6103/SHARE.w3.700
J.S. Muller et al.
1030
Bryan, M. L., & Jenkins, S. P. (2016). Multilevel modelling of country effects: A cautionary tale. European
Sociological Review, 32,322.
Budig, M. J., & England, P. (2001). The wage penalty for motherhood. American Sociological Review, 66,
204225.
Budig, M. J., Misra, J., & Boeckmann, I. (2012). The motherhood penalty in cross-national perspective: The
importance of work-family policies and cultural attitudes. Social Politics, 19, 163193.
Budig, M. J., Misra, J., & Boeckmann, I. (2016). Work-family policy trade-offs for mothers? Unpacking the
cross-national variation in motherhood earnings penalties. Work and Occupations, 43,119177.
Cameron, A. C., & Miller, D. L. (2015). A practitioners guide to cluster-robust inference. Journal of Human
Resources, 50,317372.
Cherlin, A. J. (2010). Demographic trends in the United States: A review of research in the 2000s. Journal of
Marriage and Family, 72,403419.
Christopher, K. (2005). A pauperization of motherhood?Singlemotherhoodandwomens poverty over
time. Journal of Poverty, 9(3), 123.
Cornwell, B. (2015). Social sequence analysis. Cambridge, MA: Cambridge University Press.
Correll, S. J., Benard, S., & Paik, I. (2007). Getting a job: Is there a motherhood penalty? American Journal of
Sociology, 112, 12971339.
Cukrowska-Torzewska, E. (2017). Cross-country evidence on motherhood employment and wage gaps: The
role of work-family policies and their interaction. Social Politics, 24,178220.
Dannefer, D. (2003). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social
science theory. Journal of Gerontology, Series B: Psychological Sciences and Social Sciences, S327
S337.
Davies, H., & Joshi, H. (1994). The foregone earnings of Europes mothers. In O. Eckert (Ed.), Standards of
living and families: Observation and analysis (pp. 101134). Paris, France: INED John Libbey Eurotext.
DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advantage as a mechanism for inequality: A review of
theoretical and empirical developments. Annual Review of Sociology, 32,271297.
Elder, G. H., Johnson, M. K., & Crosnoe, R. (2004). The emergence and development of life course theory. In
J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of the life course (pp. 319). New York, NY: Springer.
Elzinga, C. H., & Liefbroer, A. C. (2007). De-standardization of family-life trajectories of young adults: A
cross-national comparison using sequence analysis. European Journal of Population, 23, 225250.
Esping-Andersen, G. (1999). Social foundations of postindustrial economies. New York, NY: Oxford
University Press.
Esping-Andersen, G. (2009). The incomplete revolution: Adapting to womensnewroles. Cambridge, MA:
Polity Press.
Eurostat. (n.d.). ECU/EUR exchange rates versus national currencies. Retrieved from http://data.europa.eu/88
u/dataset/dopKjbZLZsXQGLTjpIkC2w
Festy, P. (1980). On the new context of marriage in western Europe. Population and Development Review, 6,
311315.
Fokkema, T., Kveder, A., Hiekel, N., Emery, T., & Liefbroer, A. (2016). Generations and Gender Programme
Wave 1 data collection: An overview and assessment of sampling and fieldwork methods, weighting
procedures, and cross-sectional representativeness. Demographic Research, 34,499524. https://doi.
org/10.4054/DemRes.2016.34.18
Fokkema, T., & Liefbroer, A. C. (2008). Trends in living arrangements in Europe: Convergence or diver-
gence? Demographic Research, 19, 13511418. https://doi.org/10.4054/DemRes.2008.19.36
Fortin, N. M. (2005). Gender role attitudes and the labour-market outcomes of women across OECD
countries. Oxford Review of Economic Policy, 21,416438.
Gabadinho, A., Ritschard, G., Müller, N. S., & Studer, M. (2011). Analyzing and visualizing state sequences
in R with TraMineR. Journal of Statistical Software, 40(4), 137.
Gangl, M. (2004). Welfare states and the scar effects of unemployment: A comparative analysis of the United
States and West Germany. American Journal of Sociology, 109, 13191364.
Gangl, M., & Ziefle, A. (2009). Motherhood, labor force behavior, and womens careers: An empirical
assessment of the wage penalty for motherhood in Britain, Germany, and the United States. Demography,
46,341369.
Gauthier,A.H.(1996).The state and the family: A comparative analysis of family policies in industrialized
countries. Oxford, MA: Clarendon Press.
Generations and Gender Programme. (2016). Generations and Gender Contextual Database.TheHague:
Netherlands Interdisciplinary Demographic Institute [Distributor]. Retrieved from https://www.ggp-i.
org/data/ggp-contextual-database/
The Long-Term Costs of Family Trajectories 1031
Goldin, C. (2006). The quiet revolution that transformed womens employment, education, and family.
American Economic Review: Papers & Proceedings, 96,121.
Gornick, J. C., & Meyers, M. K. (2003). Families that work: Policies for reconciling parenthood and
employment. New York, NY: Russell Sage Foundation.
Gornick, J. C., Meyers, M. K., & Ross, K. E. (1997). Supporting the employment of mothers: Policy variation
across fourteen welfare states. Journal of European Social Policy, 7,4570.
Gough, M. (2017). Birth spacing, human capital, and the motherhood penalty at midlife in the United States.
Demographic Research, 37, 363416. https://doi.org/10.4054/DemRes.2017.37.13
Gough, M., & Noonan, M. (2013). A review of the motherhood wage penalty in the United States. Sociology
Compass, 7, 328342.
Hakim, C. (2000). Work-lifestyle choices in the 21st century: Preference theory. New York, NY: Oxford
University Press.
Hakim, C. (2003). A new approach to explaining fertility patterns: Preference theory. Population and
Development Review, 29,349374.
Hakim, C. (2004). Key issues in womens work: Female diversity and the polarisation of womens employment
(2nd ed.). New York, NY: Routledge.
Halldén, K., Levanon, A., & Kricheli-Katz, T. (2016). Does the motherhood wage penalty differ by individual
skill and country family policy? A longitudinal study of ten European countries. Social Politics, 23,363
388.
Han, S. Y., Liefbroer, A. C., & Elzinga, C. H. (2017). Comparing methods of classifying life courses:
Sequence analysis and latent class analysis. Longitudinal and Life Course Studies, 8,319341.
Hantrais, L., & Letablier, M.-T. (1996). Families and family policies in Europe. New York, NY: Longman.
Harkness, S., & Waldfogel, J. (2003). The family gap in pay: Evidence from seven industrialized countries.
Research in Labor Economics, 22 369414.
Härkönen, J., & Dronkers, J. (2006). Stability and change in the educational gradient of divorce: A comparison
of seventeen countries. European Sociological Review, 22,501517.
Heinz, W. R., Huinink, J., Swader, C. S., & Weymann, A. (Eds.). (2009). The life course reader: Individuals
and societies across time. Frankfurt, Germany: Campus Verlag.
Hiekel, N., Liefbroer, A. C., & Poortman, A.-R. (2014). Understanding diversity in the meaning of cohab-
itation across Europe. European Journal of Population, 30,391410.
Huber, P. J. (1967). The behavior of maximum likelihood estimates under nonstandard conditions. In
Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability (Vol.1,pp.
221233). Berkeley: University of California Press.
Jalovaara, M., & Fasang, A. (2019). Family life courses, gender, and mid-life earnings. European Sociological
Review. https://doi.org/10.1093/esr/jcz057
Juhn, C., & Mccue, K. (2017). Specialization then and now: Marriage, children, and the gender earnings gap
across cohorts. Journal of Economic Perspectives, 31(1), 183204.
Kiernan, K. (2004). Unmarried cohabitation and parenthood in Britain and Europe. Law & Policy, 26,33
55.
https://do i.o rg/10.1111/j.0265-8240.2004.00162.x
Killewald, A., & García-Manglano, J. (2016). Tethered lives: A couple-based perspective on the consequences
of parenthood for time use, occupation, and wages. Social Science Research, 60,266282.
Killewald, A., & Gough, M. (2013). Does specialization explain marriage penalties and premiums? American
Sociological Review, 78, 477502.
Killewald, A., & Zhuo, X. (2015). Motherslong-term employment patterns (Upjohn Institute Working Paper
No. 15-247). Kalamazoo, MI: Upjohn Institute for Employment Research.
Langner, L. A. (2015). Within-couple specialisation in paid work: A long-term pattern? A dual trajectory
approach to linking lives. Advances in Life Course Research, 24,4765.
Mandel, H., & Semyonov, M. (2006). A welfare state paradox: State interventions and womens employment
opportunities in 22 countries. American Journal of Sociology, 111, 19101949.
Miller, A. R. (2011). The effects of motherhood timing on career path. Journal of Population Economics, 24,
10711100.
Mincer, J. (1975). Education, experience, and the distribution of earnings and employment: An overview. In F.
T. Juster (Ed.), Education, income, and human behavior (pp. 7194). New York, NY: McGraw-Hill.
Mincer, J., & Polachek, S. (1974). Family investments in human capital: Earnings of women. Journal of
Political Economy, 82(2, part 2), S76S108.
Misra, J., Moller, S., & Budig, M. J. (2007). Work-family policies and poverty for partnered and single women
in Europe and North America. Gender & Society, 21,804827.
Moen, P., Robinson, J., & Fields, V. (1994). Womens work and caregiving roles: A life course approach.
Journal of Gerontology: Social Sciences, 49,S176S186.
J.S. Muller et al.
1032
Moilanen, S., May, V., Räikkönen, E., Sevón, E., & Laakso, M.-L. (2016). Mothersnon-standard working
and childcare-related challenges: A comparison between lone and coupled mothers. International Journal
of Sociology and Social Policy, 36(1/2), 3652.
Muller,J.S.(2016).Preparation of the earnings variable in the GGS data Wave 1 (Working paper). The
Hague: Netherlands Interdisciplinary Demographic Institute (NIDI).
Murphy, K. M., & Welch, F. (1990). Empirical age-earnings profiles. Journal of Labor Economics, 8,202
229.
Myrdal,A.,&Klein,V.(1956).Wom e n s two roles: Home and work. London, UK: Routledge.
Nieuwenhuis, R., & Maldonado, L. C. (2018). The triple bind of single-parent families: Resources, employ-
ment and policies. In R. Nieuwenhuis & L. C. Maldonado (Eds.), The triple bind of single-parent
families: Resources, employment and policies to improve wellbeing (pp. 130). Bristol, UK: Policy Press.
ORand, A. M., & Henretta, J. C. (1982). Delayed career entry, industrial pension structure, and early
retirement in a cohort of unmarried women. American Sociological Review, 47, 365373.
Orloff, A. S. (1993). Gender and the social rights of citizenship: The comparative analysis of gender relations
and welfare states. American Sociological Review, 58, 303328.
Pascall, G., & Manning, N. (2000). Gender and social policy: Comparing welfare states in central and eastern
Europe and the former Soviet Union. Journal of European Social Policy, 10,240266.
Perelli-Harris, B., Kreyenfeld, M., & Kubisch, K. (2015). Harmonized histories: Manual for the preparation
of comparative fertility and union histories (MPIDR Working Paper 2010-011, as part of the Nonmarital
Childbearing Project). Rostock, Germany: Max Planck Institute for Demographic Research.
Pienta, A. (1999). Early childbearing patterns and womens labor force behavior in later life. Journal of
Wom e n & A g i n g, 11(1), 6984.
Polanchek, S. W. (1995). Earnings over the life cycle: What do human capital models explain? Scottish
Journal of Political Economy, 42,267289.
Prince Cooke, L. (2011). Gender-class equality in political economies.NewYork,NY:Routledge.
Raymo, J. M., Carlson, M., VanOrman, A., Lim, S.-J., Perelli-Harris, B., & Iwasawa, M. (2015). Educational
differences in early childbearing: A cross-national comparative study. Demographic Research, 33,6592.
https://doi.org/10.4054/DemRes.2015.33.3
Rogers, W. H. (1993). Regression standard errors in clustered samples. In S. Becketti (Ed.), Stata technical
bulletin (Vol. 13, pp. 1923). College Station, TX: Stata Press.
Roman, C. (2017). Between money and love: Work-family conflict among Swedish low-income single
mothers. Nordic Journal of Working Life Studies, 7(3), 2341. https://doi.org/10.18291/njwls.v7i3.97093
Saraceno, C. (2011). Childcare needs and childcare policies: A multidimensional issue. Current Sociology, 59,
7896.
Saraceno, C., & Keck, W. (2010). Can we identify intergenerational policy regimes in Europe? European
Societies, 12, 675696.
Saraceno, C., & Keck, W. (2011). Towards an integrated approach for the analysis of gender equity in policies
supporting paid work and care responsibilities. Demographic Research, 25, 371406. https://doi.
org/10.4054/DemRes.2011.25.11
Schröder, M. (Ed.). (2011). Retrospective data collection in the Survey of Health, Ageing and Retirement in
Europe: SHARELIFE methodology. Mannheim, Germany: Mannheim Research Institute for the
Economics of Aging (MEA).
Sigle-Rushton, W., & Waldfogel, J. (2007a). The incomes of families with children: A cross-national
comparison. Journal of European Social Policy, 17,299318.
Sigle-Rushton, W., & Waldfogel, J. (2007b). Motherhood and womens earnings in Anglo-American,
continental European, and Nordic countries. Feminist Economics, 13(2), 5591.
Snijders, T. A. B., & Bosker, R. J. (2012). Multilevel analysis: An introduction to basic and advanced
multilevel modeling (2nd ed.). Thousand Oaks, CA: Sage.
Sobotka, T. (2004). Postponement of childbearing and low fertility in Europe. Amsterdam, the Netherlands:
Dutch University Press.
Sobotka, T., & Toulemon, L. (2008). Overview chapter 4: Changing family and partnership behaviour:
Common trends and persistent diversity across Europe. Demographic Research, 19,85138.
https://doi.org/10.4054/DemRes.2008.19.6
Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87, 355374.
Studer, M. (2013). WeightedCluster library manual: A practical guide to creating typologies of trajectories in
the social sciences with R (LIVES Working Paper No. 24).Geneva, Switzerland: Swiss National Centre
of Competence in Research.
Studer, M., & Ritschard, G. (2014). A comparative review of sequence dissimilarity measures (LIVES
Working Paper No. 33).Geneva, Switzerland: Swiss National Centre of Competence in Research.
The Long-Term Costs of Family Trajectories 1033
Teachman, J. D., Tedrow, L. M., & Crowder, K. D. (2000). The changing demography of Americas families.
Journal of Marriage and the Family, 62,12341246.
Tijdens, K. (1993). Organisatie en Financiering van Kinderopvang [Organization and financing of childcare].
In K. Tijdens & S. Lieon (Eds.), Kinderopvang in Nederland: Organisatie en financiering financiering
[Childcare in The Netherlands: Organization and financing] (pp. 79). Utrecht, the Netherlands:
Uitgeverij Jan van Arkel.
University of Essex Institute for Social and Economic Research. (2010). British Household Panel Survey:
Wav e s 1 18, 19912009 (7th ed.) [Data set]. Colchester: UK Data Service. https://doi.org/10.5255
/UKDA-SN-5151-1
Verbakel, E., Luijkx, R., & de Graaf, P. M. (2008). The association between husbandsand wiveslabor
market positions in the Netherlands. Research in Social Stratification and Mobility, 26, 257276.
Vikat, A., Spéder, Z., Beets, G., Billari, F., Bühler, C., Fokkema, T., . . . Solaz, A (2007). Generations and
Gender Survey (GGS): Towards a better understanding of relationships and processes in the life course.
Demographic Research, 17, 389440. https://doi.org/10.4054/DemRes.2007.17.14
Warren, J. R., Luo, L., Halpern-Manners, A., Raymo, J. M., & Palloni, A. (2015). Do different methods for
modeling age-graded trajectories yield consistent and valid results? American Journal of Sociology, 120,
18091856.
White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for
heteroskedasticity. Econometrica, 48, 817830.
World Bank. (2013). Measuring the real size of the world economy: The framework, methodology, and results
of the international comparison programICP. Washington, DC: World Bank. https://doi.org/10.1596
/978-0-8213-9728-2
World Bank. (n.d.-aa). Consumer price index (2010 = 100). Retrieved from https://data.worldbank.
org/indicator/FP.CPI.TOTL
World Bank. (n.d.-bb). PPP conversion factor, private consumption (LCU per international$). Retrieved from
https://data.worldbank.org/indicator/PA.NUS.PRVT.PP
Publishers Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
Affiliations
Joanne S. Muller
1,2
&Nicole Hiekel
1,3
&Aart C. Liefbroer
1,4,5
1
Netherlands Interdisciplinary Demographic Institute (NIDI), PO Box 11650, 2502 AR The Hague,
The Netherlands
2
University of Groningen, PO Box 72, 9700 AB Groningen, The Netherlands
3
Department of Sociology and Social Psychology, University of Cologne, Albertus-Magnus-Platz,
50923 Cologne, Germany
4
Department of Epidemiology, University Medical Center Groningen (UMCG), University of Groningen,
PO Box 30001, 9700 RB Groningen, The Netherlands
5
Department of Sociology, Vrije Universiteit Amsterdam, De Boelelaan 1081, 1081 HV Amsterdam,
The Netherlands
J.S. Muller et al.
1034
... We follow Henz (2004) life course focus and add to it by theorizing and empirically testing the consequences of informal care for wages and in doing so take a closer look at long-term rather than short-term consequences. The literature on wage penalties due to caregiving from a life course perspective mostly focuses on childcare (e.g., Muller, Hiekel, & Liefbroer, 2020). Informal care, however, is in general far less predictable than childcare and it can intersect with paid work at more various moments throughout the life course (Ehrlich, Möhring, & Drobnič, 2019;Henz, 2004). ...
... Furthermore, disadvantages in the domain of paid work accumulate over the life span: inequalities earlier in the employment career intensify and the loss in wages will add up over the life course (Crystal, Shea, & Reyes, 2016;Möhring, 2018). A loss or stagnation in human capital will have long-lasting consequences if it occurs early in working life when it is important to gain experience, educate yourself, develop skills, show productivity, and establish a career (Abendroth et al., 2014;Florian, 2018;Muller et al., 2020). It is more influential for the employment development when the conflict occurs earlier in life as young caregivers might be perceived as less committed to work, less productive, and less suited for promotion (Abendroth et al., 2014;Ehrlich et al., 2019). ...
Article
Full-text available
Caring for a friend or family member in need of care has been found to have negative consequences for wages. This study contributes to the literature by studying how three major life course factors, namely timing of first caregiving, duration of caregiving, and the number of caregiving episodes, help to explain the (hourly) wage penalty for informal caregivers (i.e., providers of health-related care to older or disabled people in the personal network). We used unique retrospective data of 1,417 informal caregivers in the Netherlands that map start and end dates of up to seven caregiving episodes. Findings showed that a higher number of caregiving episodes was related to a stronger wage penalty, whereas timing of first caregiving was not associated with a wage penalty. Opposite to our expectation, we found that the wage penalty decreased the longer someone cared, potentially even resulting in a wage premium for long-time caregivers. We conclude that applying a life course perspective is relevant when examining employment consequences of informal caregiving and that caregiving possibly fosters skills that are beneficial for employment careers in the long run.
... Higher rates of female employment are central to the promotion of economic growth in the current context of an ageing and shrinking workforce, and to women's economic independence (Nieuwenhuis et al., 2020). Employment breaks and reduced work commitment during the childrearing years might affect mothers' earnings, future employability and income in later life (Adams et al., 2016;Grimshaw and Rubery, 2015;Muller et al., 2020). Despite policy consensus on the need to promote and support higher levels of maternal employment, the motherhood gap in labour force participation remains common across Western economies (Eurofound, 2016), and the complexity of the mechanisms at force is still to be fully understood. ...
Article
Full-text available
This article studies mothers’ employment transitions around childbirth. It argues that leaving employment around childbirth and returning after an interruption might depend on multiple influences: the micro-context of individual and household characteristics, the meso-context of women’s jobs and the macro-context of broader cultural and institutional factors. This conceptual model is tested using data from the Italian Institute of Statistics (ISTAT) ‘Family and Social Subjects (2009)’ survey. The findings confirm that mothers’ transitions out of employment are shaped by micro-characteristics such as education, meso-characteristics such as status and security of prior jobs, and macro geographical and temporal factors. Subsequent returns to employment also reflect micro and macro influences, as mothers born before 1950, with low education, and large families are less likely to return; but they seem less dependent upon prior job characteristics. The research highlights the importance of considering multiple levels of influence to understand the enabling factors of maternal employment.
... Our approach enables us to explicitly acknowledge that an aggregate of time-dependent processes featuring the occurrences, timings, and ordering of family transitions are (directly or indirectly) related to the life-long accumulation of economic resources and thus contributes to intra-cohort wealth inequality. While similar longitudinal approaches were taken by Madero-Cabib and Fasang (2016), Muller et al. (2020) or Jalovaara and Fasang (2020) to examine the association between family life courses and income in mid-to late-life, it needs to acknowledged that wealth is not a direct function of income, as aspects such as consumption or financial transfers additionally influence wealth accumulation . ...
Article
Full-text available
Considering soaring wealth inequalities in older age, this research addresses the relationship between family life courses and widening wealth differences between individuals as they age. We holistically examine how childbearing and marital histories are associated with personal wealth at ages 50-59 for Western Germans born between 1943 and 1967. We propose that deviations from culturally and institutionally-supported family patterns, or the stratified access to them, associate with differential wealth accumulation over time and can explain wealth inequalities at older ages. Using longitudinal data from the German Socio-Economic Panel Study (SOEP, v34, waves 2002-2017), we first identified typical family trajectory patterns between ages 16 and 50 with multichannel sequence analysis and cluster analysis. We then modelled personal wealth ranks at ages 50-59 as a function of family patterns. Results showed that deviations from the standard family pattern (i.e. stable marriage with, on average, two children) were mostly associated with lower wealth ranks at older age, controlling for childhood characteristics that partly predict selection into family patterns and baseline wealth. We found higher wealth penalties for greater deviation and lower penalties for moderate deviation from the standard family pattern. Addressing entire family trajectories, our research extended and nuanced our knowledge of the role of earlier family behaviour for later economic wellbeing. By using personal-level rather than household-level wealth data, we were able to identify substantial gender differences in the study associations. Our research also recognised the importance of combining marital and childbearing histories to assess wealth inequalities. Supplementary information: The online version contains supplementary material available at 10.1007/s10680-021-09601-4.
... According to various studies (Eagly et al., 2003;Van Engen and Willemsen, 2004;Antón, 2006;Eagly and Carli, 2007) one of the main barriers arises from the duality of family vs. work. In this way, the reality is that many women perceive that they must choose between home and work, for which reason they tend to abandon their development and progress in the organization, due to the incompatibility of adequately fulfilling the roles of being a good professional and being a good mother at the same time (Muller et al., 2020). ...
Article
Full-text available
The impediments and barriers that women face in entering and developing a police career have received relatively little attention from researchers. As of today in Europe, despite the slow progress, the 25% barrier to female representation has already been overcome in several countries. However, many areas remain closed to women within police organizations. In this context, research was conducted based on a content analysis of the perceptions of 56 police officers, 28 men and 28 women with considerable police experience, occupying executive leadership positions from a total of 26 European countries. Data was collected through a questionnaire composed of 23 open questions. The results show a considerable gap between the perceptions of male and female police executive leaders with regard to access, career development and workplace conditions faced by policepersons. According to the results, the mirage of equality, dominant in the view of male police officers, is a major barrier to achieving real equality, both horizontally and vertically. What implications these results have on the strategies that police organizations should follow to achieve the challenge of inclusion are discussed, and new ways of analysis are proposed.
... The fact that motherhood penalties persisted into women's midlife may reflect women's difficulties reconciling work with family under Japan's patriarchal social norms and labor market practices (Brinton, 2001;Brinton & Oh, 2019). This finding also resonates with previous studies (Kleven, Landais, Posch et al., 2019;Muller, Hiekel, & Liefbroer, 2020), showing that motherhood earnings penalty was more likely to persist to midlife for women who followed conservative gender norms and family trajectories in societies with limited work-family supports. Because younger generations in Japan remain relatively conservative in their attitudes on gender divisions of labor and female employment (Piotrowski et al., 2019), the significant motherhood earnings penalty in Japan is likely to persist in the near future. ...
Article
The issue of motherhood earnings penalty has been well-documented in many Western countries. However, only a few studies discussed how earnings penalty evolves over time and varies across different parity of birth. Moreover, related research in non-Western developed countries is scant. This study contributes to the motherhood penalty literature by examining the long-term impacts (up to 10 years after childbirth) of the first and the second birth on women’s employment, work hours, wage rates, and earnings in Japan. It proposes a novel research design based on the event-study approach and fixed effects regressions to quantify the dynamic effects of motherhood resulted from two consecutive birth transitions. Drawing on longitudinal data from the Japanese Panel Survey of Consumers (from 1993 to 2015), our results show that both the first and the second birth trigger short-term earnings penalties by causing a considerable employment slump upon pregnancy. In the long run, while women’s employment rates recover, work hours and wage rates remain significantly lower than their pre-pregnancy level, leading to the long-term earnings penalty. More importantly, the long-term negative impacts of childbirth on labor supply and wage rates result mostly from women’s first-time rather than the second-time birth transition in Japan. These findings imply that motherhood in Japan imposes long-term penalties on women’s earnings, primarily by depressing maternal labor supply after their first-time motherhood transition.
... Employment rates among single parents are high (Nieuwenhuis and Maldonado 2018a). Indeed, a recent study found that women living with dependent children and a partner are more likely to reduce their employment than women living with a partner and no children or single mothers (Muller et al. 2020). These findings challenge behavioural explanations based on the supposed joblessness of single parents. ...
Chapter
Full-text available
The objective of this chapter is to explore family poverty and how it is mitigated by social policy in European countries. It examines the poverty risks of families in Europe, with an emphasis on families with dependent children, as these families tend to be at greater risk of poverty than others because they have more members to maintain with a given income.
Article
Despite decades of progress toward gender equality, women remain as the United States’ primary caregivers. Past research has shown how couples and families organize care at distinct life course moments but has not studied how these moments combine to create differences in men and women’s full life courses of caregiving. In this article, I look beyond negotiations within households to introduce a complementary demographic explanation for the gender gap in caregiving—women’s greater likelihood to reside with dependents. A focus on patterns of coresidence is warranted, given the growing diversity of family forms, which may expose women to additional and varied care demands at differing ages. Drawing on data from the 2011 to 2019 American Time Use Surveys, I study how coresidential care demands shape the population gender gap in childcare and eldercare across ages 20–79 and how demands differ for Black, White, and Latina/o women and men. My results show that coresidence with dependents is uneven across the life course, and women’s exposures occur early and late in adulthood, while men are exposed to more care demands in midlife. Patterns of childbearing, partnership, and extended family embeddedness contribute to Black and Latina women’s greater exposure to care demands early in adulthood and White women’s greater exposure to care demands later in the life course. Thus, despite growing egalitarianism within households, the rise of complex families contributes to bolstering population-level gender inequality in caregiving across adulthood.
Article
We examine how the life courses of couples in East and West Germany are linked to women’s income in later life using multichannel sequence analysis. By applying a couple perspective, we overcome the individualistic approach in most previous research analysing women’s old-age income. Detailed monthly information on spouses’ employment and earnings trajectories from age 20 to 50 for the birth cohorts 1925–1965 (N = 2020) stems from SHARE-RV, a data linkage of the administrative records of the German public pension insurance with the Survey of Health, Ageing and Retirement in Europe (SHARE). We identify seven clusters of couples’ life courses and link them to women’s absolute individual and relative household income in later life using a cohort comparison to identify trends over time. While in older cohorts, women in male-breadwinner type clusters achieve the lowest, and those in dual-earner type couples have the highest incomes, this relationship does no longer prevail in younger cohorts. Here, we identify a polarization in dual-earner and male-breadwinner type clusters. The former increasingly diverge into successful female-breadwinner constellations and those with both partners in marginalized careers. The latter polarize between persistent male-breadwinner constellations and those in which women increase their labor market engagement.
Article
This paper identifies typical generational structures of families over individual life courses (generational placement trajectories) and investigates changes in their prevalence across cohorts. The trajectories represent changing opportunity structures for intergenerational support, and they are inherently dynamic. As of yet, they have not been analysed holistically. With data from the German Ageing Survey (https://www.dza.de/en/research/deas), for men and women of the birth cohorts 1939–43, 1944–48, and 1949–53 (n=2,607), the occurrence, timing, and ordering of generational placements defined through the (simultaneous) presence of biological parents, children, and grandchildren up to the age of 60 is examined using sequence and cluster analyses. Cohort, gender, and regional differences are examined with multinomial logistic regressions. Six different typical patterns of individual trajectories were identified, ranging from clusters with multiple transitions early in life to more stable clusters with fewer transitions later on. The most common were three-generation families, while patterns without children were the most uncommon. Across cohorts, generational structures of families changed towards “more stable” trajectories. Clusters characterised by early (grand)parenthood were more likely experienced by women and individuals growing up in former East Germany. The presented patterns of individual trajectories reveal novel detailed insights into how the opportunity structures for intergenerational family support differ between individuals and develop across individual age and socio-historical context.
Book
Full-text available
This open access book examines how childhood social disadvantage influences young-adult demographic decision-making and later-life economic and well-being outcomes. This book in particular focuses on testing whether the consequences of childhood social disadvantage for adult outcomes differ across societies, and whether these differences are shaped by the “context of opportunities” that societies offer to diminish the adverse impact of economic and social deprivation. The book integrates a longitudinal approach and provides new insights in how the experience of childhood disadvantage (e.g. low parental socio-economic status, family disruption) influences demographic decisions in adulthood (e.g. the timing of family-events such as cohabitation, marriage or parenthood; the risk of divorce or having a child outside a partner relationship; the exposure to later-life loneliness, poor health, and economic adversity). Moreover, using a cross-national comparative perspective it investigates whether the relationships of interest differ across nations, and tests the “context of opportunities” hypothesis arguing that the links between childhood disadvantage and adult outcomes are weakened in societal contexts offering good opportunities for people to escape situations of deprivation. To do so, the book analyzes national contexts based on economic prosperity, family values and norms, and welfare-state arrangements.
Article
Full-text available
There is a long-standing debate on whether extensive Nordic family policies have the intended equalizing effect on family and gender differences in economic outcomes. This article compares how the combination of family events across the life course is associated with annual and accumulated earnings at mid-life for men and women in an egalitarian Nordic welfare state. Based on Finnish register data (N = 12,951), we identify seven typical family life courses from ages 18 to 39 and link them to mid-life earnings using sequence and cluster analysis and regression methods. Earnings are highest for the most normative family life courses that combine stable marriage with two or more children for men and women. Mid-life earnings are lowest for unpartnered mothers and never-partnered childless men. Earnings gaps by family lives are small among women but sizeable among men. Gender disparities in earnings are remarkably high, particularly between men and women with normative family lives. These gaps between married mothers and married fathers remain invisible when looking only at motherhood penalties. Results further highlight a large group of (almost) never-partnered childless men with low earnings who went largely unnoticed in previous research.
Book
Full-text available
Single parents face a triple bind of inadequate resources, employment, and policies, which in combination further complicate their lives. This book - multi-disciplinary and comparative in design - shows evidence from over 40 countries, along with detailed case studies of Sweden, Iceland, Scotland, and the UK. It covers aspects of well-being that include poverty, good quality jobs, the middle class, wealth, health, children’s development and performance in school, and reflects on social justice. Leading international scholars challenge our current understanding of what works and draw policy lessons on how to improve the well-being of single parents and their children.
Article
Full-text available
We compare life course typology solutions generated by sequence analysis (SA) and latent class analysis (LCA). First, we construct an analytic protocol to arrive at typology solutions for both methodologies and present methods to compare the empirical quality of alternative typologies. We apply this protocol to develop and compare SA- and LCA-derived family-life typologies for women born between 1960 and 1964 in 15 European countries, using data from the Family and Fertility Survey. This paper contributes to the use of these classification techniques in four different ways. First, we present guidelines on how to establish the number of classes or clusters to use. Second, we show how to evaluate the stability of these clusters. Third, we provide a way to evaluate the validity of these clusters and finally, we provide for a formal heuristic to relate the stochastically defined latent classes to the distance-based clusters found with SA. © 2017, Society for Longitudinal and Life Course Studies. All rights reserved.
Article
Full-text available
Research suggests that single mothers are an increasingly vulnerable group in Scandinavia. Drawing on in-depth interviews, this study takes a closer look at challenges faced by Swedish working-class and low-income single mothers. It explores how working conditions and economic resources influence their access to valued practices, such as the possibility to reconcile paid work and family commitments. It is assumed that discrepancies between mothers’ notions of good mothering and their ability to act in accordance with these values give rise to conflicts and dilemmas. Findings show that lack of financial resources and low control over their work situation significantly limited the mothers’ possibility to combine various responsibilities and to practice the kind of mothering they preferred. Furthermore, the opening hours of preschools frequently did not match the mothers’ working schedules, and they could often not effectively benefit from some of the social rights granted by the Swedish welfare state.
Article
Full-text available
BACKGROUND Researchers have examined how first-birth timing is related to motherhood wage penalties, but research that examines birth spacing is lacking. Furthermore, little research has examined the persistence of penalties across the life course. OBJECTIVE The objective is to estimate the effects of birth spacing on midlife labor market outcomes and assess the extent to which these effects vary by education and age at first birth. METHODS I use data from the United States from the 1979-2010 waves of the National Longitudinal Survey of Youth 1979 and dynamic inverse probability of treatment weighting to estimate the effects of different birth intervals on mothers' midlife cumulative work hours, cumulative earnings, and hourly wages. I examine how education and age at first birth moderate these effects. RESULTS Women with birth intervals longer than two years but no longer than six years have the smallest penalties for cumulative outcomes; in models interacting the birth interval with age at first birth, postponement of a first birth to at least age 30 appears to be more important for cumulative outcomes than birth spacing. College-educated women benefit more from a longer birth interval than less educated women. CONCLUSIONS Childbearing strategies that result in greater accumulation of human capital provide long-run labor market benefits to mothers, and results suggest that different birthspacing patterns could play a small role in facilitating this accumulation, as theorized in past literature.
Article
Full-text available
This article investigates the effects of work–family policies on motherhood employment and wage gaps. We study twenty-eight European countries, including Central and Eastern European countries. We show that unless the leave period is exceptionally long, enhanced childcare and leave policies help women combine family responsibilities with paid work. The positive effect of leaves is, however, much weaker in countries where childcare is limited. In these countries, improving access to childcare is therefore a more efficient policy than offering extended leaves. Because extended leaves encourage women to stay outside of the labor market, they are found to lead to wage penalties.